PDDL generator for planning tasks in tourism

The Planner Test Tool is a Java-based application that generates PDDL (Planning Domain Definition Language) domain and problem files for planning tasks in the tourism domain. It provides a user-friendly interface for defining the planning task requirements and generates the necessary PDDL files.

Requirements

• Java JDK 21
• Maven
• Docker
• NetBeans 20+

Getting Started

1. Clone the repository.
2. Install Java JDK 21.
3. Install Maven.
4. Install Docker.
5. Build the project with Maven: mvn clean package.
6. Run the program using the UI or the command line.

Usage

Generation of domain and problem PDDL files using the UI

To run the program with the GUI, simply execute the following command if you are using maven:

mvn clean javafx:run

If you are using the compiled jar file, execute the following command:

java -jar target/PlannerTests-2.0.jar

From the UI customize the options, select the target folder for generated PDDL files and click on Genera PDDL.

Generation of domain and problem PDDL files using the command line

To run the program from the command line, you can use the following options:

• -i or --input: Input museum.properties file
• -o or --output: Directory where to save the PDDL files (required)
• -r or --rooms: Number of rooms
• -a or --attractions: Number of attractions
• -l or --links: Number of links between rooms (must be between 1 and 3)
• -v or --visits: Number of attractions to visit (required)

You must specify either an input museum properties file or provide the number of rooms, attractions, and links.

Example Command

If you are using maven, you can run the program with the following command:

mvn clean compile exec:java -Dexec.args="-o output_directory -r 10 -a 50 -l 3 -v 20"

If you are using the compiled jar file, you can run the program with the following command:

java -jar target/PlannerTests-2.0.jar -o output_directory -r 10 -a 50 -l 5 -v 20

This command will generate PDDL files with the specified number of rooms, attractions, links, and visits, and save them in the specified output directory.

Using an Input File

If you have an input museum.properties file, you can specify it with the -i option:

mvn clean compile exec:java -Dexec.args="-i path/to/museum.properties -o output_directory -v visits"

or

java -jar target/PlannerTests-2.0.jar -i path/to/museum.properties -o output_directory -v visits

In this case, the program will use the information from the input file to generate the PDDL files.

Output

The program will generate the following files in the specified output directory:

• domain.pddl: The PDDL domain file
• problem.pddl: The PDDL problem file
• generation_info.json: A JSON file containing information about the generation process, including the number of rooms, attractions, links, visits, and the time taken to generate the files.
• museum.properties: The topology of the museum

If you want to use custom domain and problem files you must rename them to domain.pddl and problem.pddl and place them in the same folder.

At the end of any algorithm, you will find the generated output file plan.sas in the same folder.

Running the planner with Fast Downward

The following docker command gets in input the domain and problem files and generates the file plan.sas by using Fast Downward planner. The planner can be changed by using the --alias flag with the desired algorithm. The --search-time-limit flag sets the time limit for the search algorithm in seconds and is required for some algorithms, as shown in the table below. Replace <LOCAL_PDDL_FOLDER> with the directory where the PDDL files are stored and the <ALGORITHM> with one from the list below.

docker run --rm -v "<LOCAL_PDDL_FOLDER>:/pddl" aibasel/downward [--search-time-limit <TIME_LIMIT>] --plan-file /pddl/plan.sas --alias <ALGORITHM> /pddl/domain.pddl /pddl/problem.pddl

Available algorithms for the --alias flag:

Algorithm	Description	Needs --search-time-limit flag
lama	Lazy A*	No
lama-first	Lazy A* with first solution	No
seq-opt-bjolp	Sequential optimization with BJO-like pruning	No
seq-opt-fdss-1	Sequential optimization with Fast Downward single search	Yes
seq-opt-fdss-2	Sequential optimization with Fast Downward single search	Yes
seq-opt-lmcut	Sequential optimization with landmark-cut heuristic	No
seq-opt-merge-and-shrink	Sequential optimization with merge-and-shrink heuristic	Yes
seq-sat-fd-autotune-1	Sequential satisficing with Fast Downward autotuned search	No
seq-sat-fd-autotune-2	Sequential satisficing with Fast Downward autotuned search	No
seq-sat-fdss-1	Sequential satisficing with Fast Downward single search	Yes
seq-sat-fdss-2	Sequential satisficing with Fast Downward single search	Yes
seq-sat-fdss-2014	Sequential satisficing with Fast Downward single search (2014)	Yes
seq-sat-fdss-2018	Sequential satisficing with Fast Downward single search (2018)	Yes
seq-sat-lama-2011	Sequential satisficing with Lazy A* (2011)	No

Running the planner with Downward A* Blind algorithm

Use the following command to run the planner with the Downward A* Blind algorithm. Replace <LOCAL_PDDL_FOLDER> with the directory where the PDDL files are stored.

docker run --rm -v "<LOCAL_PDDL_FOLDER>:/pddl" aibasel/downward /pddl/domain.pddl /pddl/problem.pddl --search "astar(blind())"

Running the planner with SymBA* 2 algorithm

Use the following command to run the planner with the SymBA* 2 algorithm. Replace <LOCAL_PDDL_FOLDER> with the directory where the PDDL files are stored.

docker run --rm -v "<LOCAL_PDDL_FOLDER>:/pddl" ansep/symba2-64bit

Running the planner with Complementary2 algorithm

Use the following command to run the planner with the Complementary2 algorithm. Replace <LOCAL_PDDL_FOLDER> with the directory where the PDDL files are stored.

docker run --rm -v "<LOCAL_PDDL_FOLDER>:/pddl" ansep/complementary2

Running the planner with Ragnarok algorithm

Use the following command to run the planner with the Ragnarok algorithm. Replace <LOCAL_PDDL_FOLDER> with the directory where the PDDL files are stored.

docker run --rm -v "<LOCAL_PDDL_FOLDER>:/pddl" ansep/ragnarok

License

This project is licensed under the MIT License.